Dear Anastassia, dear all,
In both Life Stability and BRE you seem to use the terms strictly, tightly or rigidly correlated (in G&M 2004 also auto-correlated) as opposed to chaotic, non- or weakly-correlated.
I struggle to apply these terms. Charles Parrow in his book Normal Accidents (1984, p. 89) instead uses the engineering terms tight coupling and loose coupling. In table 3.2 (see attachment) he lists tendencies for tight as well as loose coupling. I first started looking into these terms because the tendencies of loose coupling seemed to link up well with somatic flexibility (sensu Bateson), which allows for the acclimatisation of motile animals to a range of 'outer' environmental conditions (e.g. temperature and altitude). Initially I thought the terms tight correlation and tight coupling and the terms weak correlation and loose coupling were interchangeable. However, after a closer examination this does not seem to be the case. Currently the application of Parrow’s terms seem to be more practical, but I may be wrong and as the terms are crucial to my understanding of the regulation of the local environment by immotile ecological communities and the regulation of the inner environment by motile ecological communities (host plus microbiome) I would like to discuss this.
(1) You refer to higher plants as weakly correlated. However, when I try to categorise them according to tendencies mentioned in table 3.2 they seem tightly coupled with deliberately designed-in redundancies (e.g. leaves, branches, roots). That is how I currently understand the lack of correlation between respectively leaves, branches, and roots that you point out.
(2) You refer to ecological communities as rigidly correlated. When I try to categorise them according to tendencies mentioned in table 3.2 they seem tightly coupled with deliberately designed-in buffers (i.e. the biotic regulation reservoirs from figure 3 op page 5 of G&M 2020). As you point out these communities function as rigidly correlated in particularly once the biotic sensitivity threshold has been crossed. However, if the perturbation of the local environment crosses the threshold of admissible perturbations this correlated functioning will seize (i.e. the ecological community shifts from a normal community to a decay community).
(3) You refer to individual animals as strictly correlated. When I try to categorise them according to the tendencies mentioned in table 3.2 they seem loosely coupled, which would allow for somatic flexibility (sensu Bateson) or a secondary feedback system (sensu Ashby) in which feedback is linked to feedback among its various subsystems. According to Ashby this is what enables an animal to resist systemic disruptive effects rather than only local disruptions (e.g. a cat moving away from the fire).
(4) You refer to anthills and beehives as weakly coordinated (i.e. similar to true multi-cellular organism). I am a bit confused by this comparison to multi-cellular organisms as animals, which are multi-cellular are strictly correlated. When I try to categorise them according to the tendencies mentioned in table 3.2 they seem loosely coupled, which would allow for social flexibility (sensu Bateson).
I would very much like to hear your take on this.
Best,
Arie